Proaction reports 60% sales increase after adopting Codex
Fleet-management software startup Proaction says its use of OpenAI’s Codex has increased sales by 60% and saved more than 75 hours a month across engineering and founder work. The North American company builds software for operators of cars, trucks and construction equipment, and also uses GPT‑Live‑1, GPT‑6 Astra and ChatGPT‑5.6 Sol in its platform.
Proaction’s sales process depends on showing prospective customers how its product can fit their specific vehicles and operating workflows. Before adopting Codex, the founders often relied on conversations and slide decks because tailored demonstrations required engineering time that the company could not readily allocate.
Custom demonstrations without an engineering handoff
Colin Knudsen, Proaction’s co-founder and COO, now creates four to six interactive demos each month using Codex. He says each takes 30 to 45 minutes to build. After a sales call, he directs Codex to Granola recordings, customer email threads and shared spreadsheets; Codex then creates an HTML demo environment modelled on Proaction’s product and the prospect’s fleet.
Knudsen estimates that creating equivalent demonstrations through the engineering team would require about 10 hours per demo. At four to six demos a month, Proaction puts the avoided engineering effort at 40 to 60 hours monthly. The company says the share of deals moving from initial contact into solution development rather than nurture rose by 50% to 60% when it began offering customized demos.
The demo can also become a visual reference when a prospect becomes a customer, reducing questions about what the engineering team should build. This use of Codex follows a broader pattern in which coding agents extend technical work beyond handoffs illustrates how coding agents can extend technical work beyond conventional engineering handoffs.
Connecting sales, support and product work
Knudsen uses Codex plugins for Granola, Gmail, Slack, Linear, GitHub and HubSpot to bring customer context into one workspace. He uses call transcripts and email history to prepare follow-ups, create Linear issues and update HubSpot opportunities. A scheduled automation reviews recent calls and prepares sales updates for the team.
He estimates that Codex saves 25 to 33 hours a month across 15 to 20 distinct daily tasks. Proaction has also built a customer solution center where prospects can explore workflows tailored to their business and review sales materials, giving non-engineering colleagues a way to translate customer conversations into clearer requirements.
Voice agents for fleet operations
Beyond Codex, Proaction uses ChatGPT‑5.6 Sol to help identify vehicle damage from photos submitted with issue reports. It is building what it calls a Managed Execution Layer with GPT‑Live‑1 and GPT‑6 Astra. Customers can ask specialized agents to handle work such as tolls or service, or configure workflows that assign the appropriate agent automatically.
One agent, Marty, is intended to coordinate vehicle maintenance by speaking with drivers, calling repair shops, arranging service, and helping with estimate approval and payment. Proaction says its team intervenes when human review or action is required. For businesses, the case shows the practical value of connecting customer evidence, workflow tools and human oversight before handing work to engineering or autonomous agents.